Chala O. Models, methods and information technologies for automated management of temporal knowledge bases to support management decisions

In the dissertation work, the actual scientific and applied problem of developing conceptual foundations, models, methods and information technologies for automated management of temporal knowledge bases to improve the efficiency of the choice of managerial decisions when solving partially structured and unstructured tasks is solved. For the first time, a model of a multivariate management decision with the use of temporal knowledge is proposed, which makes it possible to build and evaluate alternatives when solving the problem of forming a managerial decision. For the first time, a model is proposed for representing temporal dependencies for the tasks of supporting managerial decisions, which allows in the temporal and object aspects to adequately reflect the multivariance of the managerial decision with a given degree of detailing of temporal knowledge on the organizational hierarchy of the control object. For the first time, a model of a temporal knowledge base is proposed that contains a logical-probabilistic representation of knowledge, knowledge of the current control processes and states of the control object. For the first time, a method is proposed for identifying an anomalous state of a control object based on temporal rules, which creates conditions for automating the task of identifying a problem situation in the process of supporting managerial decisions. For the first time, a method of inverse probabilistic inference based on temporal knowledge is proposed, which makes it possible to reduce the number of alternatives from which a decision maker chooses and, thereby, to increase the efficiency of support for managerial decisions. The production representation of knowledge in the form of an ensemble of models of temporal rules has been further developed, which make it possible to use temporal knowledge to form new sequences of states of the control object, ensuring the achievement of the target state in the process of supporting managerial decisions. Methods for constructing productive knowledge in the aspect of temporal rules have been improved, which makes it possible to form temporal knowledge to support managerial decisions in automatic or automated modes. The method for determining the weights of temporal rules based on the values of the probability of realizing sequences of facts has been improved; it makes it possible to refine the weights of temporal rules after changing the state of the control object. The results of the experimental verification of the proposed methods, as well as the approbation of the developed technologies, which confirmed the operability and efficiency of the technologies for the automated management of the temporal knowledge base for solving the problems of supporting managerial decisions, are presented.